gpu: transpose patches faster
#6829Comparing
transpose-patches-fix
(c9e1e79
) withdevelop
(5d6a3c8
)+24%
- Improvement
- 3
- Untouched
- 391
- Skipped
- 2052
No successful run was found on develop ( 761c404) during the generation of this report, so 5d6a3c8 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report.
Benchmarks
2446 total
take_map[(0.1, 0.5)]
vortex-array/benches/take_patches.rs+24%
2.6 ms2.1 ms
take_map[(0.1, 1.0)]
vortex-array/benches/take_patches.rs+21%
4.2 ms3.5 ms
take_map[(0.1, 0.1)]
vortex-array/benches/take_patches.rs+11%
1,007.5 µs908.7 µs
take_map[(0.1, 0.05)]
vortex-array/benches/take_patches.rs+5%
805.9 µs766.4 µs
binary_search_std
vortex-array/benches/search_sorted.rs+4%
682.5 ns653.3 ns
binary_search_vortex
vortex-array/benches/search_sorted.rs+3%
938.9 ns909.7 ns
chunked_dict_primitive_into_canonical[f32, (1000, 100, 100)]
vortex-array/benches/chunked_dict_builder.rs+2%
961.2 µs938 µs
chunked_dict_primitive_into_canonical[f32, (1000, 10, 10)]
vortex-array/benches/chunked_dict_builder.rs+2%
124.1 µs121.3 µs
take_map[(0.01, 0.05)]
vortex-array/benches/take_patches.rs+2%
202.8 µs198.7 µs
chunked_dict_primitive_into_canonical[u64, (1000, 100, 10)]
vortex-array/benches/chunked_dict_builder.rs+1%
168.4 µs166.3 µs
chunked_opt_bool_into_canonical[(10, 1000)]
vortex-array/benches/chunk_array_builder.rs+1%
1.8 ms1.7 ms
chunked_opt_bool_into_canonical[(1000, 10)]
vortex-array/benches/chunk_array_builder.rs+1%
97.6 µs96.5 µs
encode_primitives[f32, (1000, 8)]
vortex-array/benches/dict_compress.rs+1%
136 µs134.6 µs
encode_primitives[f32, (1000, 4)]
vortex-array/benches/dict_compress.rs+1%
135 µs133.6 µs
encode_primitives[f32, (1000, 32)]
vortex-array/benches/dict_compress.rs+1%
143.1 µs141.7 µs
encode_primitives[f32, (1000, 2)]
vortex-array/benches/dict_compress.rs+1%
133.3 µs132 µs
chunked_opt_bool_into_canonical[(100, 100)]
vortex-array/benches/chunk_array_builder.rs+1%
347.2 µs344.1 µs
encode_primitives[i64, (1000, 4)]
vortex-array/benches/dict_compress.rs+1%
137.8 µs136.6 µs
chunked_dict_primitive_into_canonical[u32, (1000, 10, 10)]
vortex-array/benches/chunked_dict_builder.rs+1%
157.3 µs155.9 µs
encode_primitives[i64, (1000, 8)]
vortex-array/benches/dict_compress.rs+1%
139 µs137.8 µs
encode_primitives[f32, (10000, 4)]
vortex-array/benches/dict_compress.rs+1%
771.5 µs765.1 µs
encode_primitives[i64, (1000, 32)]
vortex-array/benches/dict_compress.rs+1%
146.7 µs145.5 µs
encode_primitives[f32, (10000, 8)]
vortex-array/benches/dict_compress.rs+1%
772.2 µs765.8 µs
chunked_dict_primitive_into_canonical[u32, (1000, 100, 100)]
vortex-array/benches/chunked_dict_builder.rs+1%
942.7 µs935 µs
encode_primitives[f32, (10000, 32)]
vortex-array/benches/dict_compress.rs+1%
781.8 µs775.4 µs
Commits
Click on a commit to change the comparison rangeBase
develop
5d6a3c8
+23.62%
gpu: transpose patches 4x faster
c9e1e79
13 hours ago
by a10y